The Structure and Classification of the Arthropoda

STRUCTURE AND CLASSIFICATION OF THE AIITHHOPODA. 523

The Structure and Classification of the Arthropoda.

By E. Ray Lankcster, M.A., LX.D., F.R.S., Director of the Natural History Departments of the British Museum.

With PUe 42.

[BY the great kindness of the proprietors of the tenth edition of the 'Encyclopaedia Britannica' I have received per- mission to reprint in this journal the articles ARTHROPODA and ARACHNIDA, which I contributed to its pages. I have been anxious that morphologists should consider the views which I have put forward in these articles (written now nearly four years ago). At the same time I have observed that they have entirely escaped the notice of two authors who have recently written general essays on the Arthropoda, viz. Dr. A. S. Packard, of Salem, Mass., and Mr. G. H. Carpenter, of Dublin. I have revised both articles only in regard to verbal inaccuracies, excepting where I have definitely stated that new matter is introduced. I hope that in their present form these articles will not fail to come under the notice of s.—E. R. L.]

ARTHROPODA is the name of one of the three sub-phyla into which one of the great phyla (or primary branches) of 524 E. EAY LANKBSTER. coeloinoccelous animals—the Appendiculata—is divided, the other two being respectively the Ch^topoda and the Rotifera. The word " Arthropoda" was first used in classification by Siebold and Stannius (' Lehrbuch der vergleich. Auatomie/ Berlin, 1845) as that of a primary division of animals, the others recognised in that treatise being Protozoa, Zoophyta, Vermes, Mollusca, and Vertebrata. The names Oondylopoda and Gnathopoda have been subsequently proposed for the same group. The word refers to the jointing of the chitinised exo-skeleton of the limbs or lateral appendages of the animals included, which are, roughly speaking, the Crustacea, Arach- nida, Hexapoda (so-called "true insects"), Centipedes, and Millipedes. This primary group was set up to indicate the residuum of Cuvier's Articnlata when his class Annelides (the modern Cheefcopoda) was removed from that " embranche- ment." At the same time Siebold and Stannius renovated the group Vermes of• Linuasus, and placed in it the ChEefcopods and the parasitic worms of Cuvier, besides the Rotifers and Turbellarian worms.1

1 As a matter of fact the group Arthropoda itself, thus constituted, was precisely identical in its area with the class Insecta of Linnaeus, the Entoma of Aristotle. But by causes which it is not easy to trace the word "Insect" had become limited since the days of Linnaeus to the Hexapod Pterygote forms, to the exclusion of his Aptera. Lamarck's penetrating genius is chiefly responsible for the shrinkage of the word Insecta, since it was lie who, forty years after Linnseus's death, set up and named the two great classes Crustacea and Araclmida (included by Linnseus under Insecta as the order "Aptera") assigning to them equal rank with the remaining Insecta of Linneeus, for which he proposed the very appropriate class-name " Hexapoda." Lamarck, however, appears not to have insisted on this name Hexapoda, and so the class of Pterygote Hexapods came to retain the group-name Insecta, which is, historically or etymological ly, no more appropriate to them than it is to the classes Crustacea and Araclmida. The tendency to retain the original name of an old and comprehensive group for one of the fragments into which such group becomes divided by the advance of knowledge—instead of keeping the name for its logical use as a comprehensive term, including the new divisions, each duly provided with a new name—is most curiously illustrated in the history of the word Physiology. Cicero says, "Physiologia naturae ratio," and such was the meaning of the name Physiologus, given to a STOUCTGBE AND CLASSIFICATION OF THE AllTHROPODA. 525 The result of the knowledge gained in the last quarter of the nineteenth century has been to discredit altogether the group Vermes, thus set up and so largely accepted by German writers even at the present day. We have, in fact, returned very nearly to Cuvier's conception of a great division or branch, which he called Avticnlata, including the Arthropoda and the Chtetopoda (the latter equivalent to the Annelides of Lamarck, a name adopted by Guvier), and differing from it only by the inclusion of the Rotif era. The name Articulata, introduced by Cuvier, has not been retained by subsequent writers. The same, or nearly the same assemblage of animals has beeu called Entomozoariaby DeBlaiuville (1882), ArthrozoabyBurmeister (1843), Entornozoa or Anuellata by Milne-Edwards (1855), and Annulosa byM'Leay (1819), who was followed by Huxley (1856). The character pointed to by all these terms is that of a ring-like segmentation of the body. This, however, is not the character to which we now ascribe the chief weight as evidence of the genetic affinity and monophyletic (uni- ancestral) origin of the Chastopods, Rotifers, and Arthropods. It is the existence in each ring of the body of a pair of hollow lateral appendages or parapodia, moved by intrinsic muscles and penetrated by blood-spaces, which is the leading fact indicating the affinities of these great sub-phyla, and uniting them as blood relations. The pai-apodia (fig. 7) of the marine branchiate worms are the same things genetically as the "legs" of Crustacea and insects (fags. 9 and 10). Hence the term Appendiculata was introduced by Lankester

cyclopaedia of wliat was kuown and imagined about earth, sea, sky, birds, beasts, and fishes, which for a thousand years was the authoritative source of information on these matters, and was translated into every European tongue. With the revival of learning, however, first one and then another special study became recognised—anatomy, botany, zoology, mineralogy, until at last the great comprehensive term Physiology was bereft of all its once-included subject-matter excepting the study of vital processes pursued by the more learned members of the medical profession. Professional tradition, and an astute perception on iheir part of the omniscience suggested by the terms, have left the medical men in English-speaking lands in undisturbed but illogical possession of the words physiology, physic, and physician. 526 E. RAY LANKESTER.

(preface to the English edition of Gegenbaur's ' Comparative Anatomy,' 1878) to indicate the group. The relationships of the Arthropoda thus stated are shown in the subjoined table : f Sub-phylum 1. Rotifera. Phylum APPENDICDLATA^ „ 2. Chsetopoda. I ,, 3. Arthropoda.

The Rotifera are characterised by the retention of what appears in Molluscs and Chaatopods as an embryonic organ, the velum or ciliated prasoral girdle, as a locomotor and food-seizing apparatus, and by the reduction of the muscular parapodia to a rudimentary or non-existent condition in all present surviving forms except Pedalion. In many important respects they are degenerate—reduced both in size and elaboration of structure. The Chastopoda are characterised by the possession of horny epidermic chsetas embedded in the integument and moved by muscles. Probably the chaatse preceded the development of parapodia, and by their concentration, and that of the muscular bundles connected with them at the sides of each segment, led directly to the evolution of the parapodia. The parapodia of Chtetopoda are never coated with dense chitin, and are, therefore, never converted into jaws; the primitive "head-lobe" or prostomium persists, and frequently carries eyes and sensory tentacles. Further, in all members of the sub-phylum Chastopoda the relative position of the prostomium, mouth, and peristomium or first ring of the body retains its primitive character. We do not find in Chastopoda that parapodia, belonging to primitively post-oral rings or body-segments (called " somites," as proposed by H. Milne-Edwards), pass in front of the mouth by adaptational shifting of the oral aperture. (See, however, 8.) The Arthropoda might be better called the " Gnathopoda/' since their distinctive character is that one or more pairs of appendages behind the mouth are densely chitinised and turned (fellow to fellow on opposite sides) towards one STBnCTUBB AND CLASSIFICATION OF THE ARTHROPODA. 527 auother so as to act as jaws. This is facilitated by an important general change in the position of the parapodia; their basal attachments are all more ventral in position than in the Chtetopoda, and tend to approach from the two sides towards the mid-ventral line. Very usually (but not in the Onychophora = Peripatus) all the parapodia are plated with chitin secreted by the epidermis, and divided into a series of joints—giving the " arthropodous" or hinged character. There are other remarkable and distinctive features of structure which hold the Arthropoda together, and render it impossible to conceive of them as having a polyphyletic origin,—that is to say, as having originated separately by two or three distinct lines of descent from lower animals; and, on the contrary, establish the view that they have been developed from a single line of primitive Gnathopods which arose by modification of parapodiate annulate worms not very unlike some of the existing ChEetopods. These additional features are the following:—(1) All existing Arthropoda have an ostiate heart and have undergone " phleboedesis," that is to say, the peripheral portions of the blood-vascular system are not fine tubes as they are in the Cheetopoda and as they were in the hypothetical ancestors of Arthropoda, but are swollen so as to obliterate to a large extent the coelom, whilst the separate veins entering the dorsal vessel or heart have coalesced, leaving valvate ostia (see Pig. 1*) by which the blood passes from a pericardial blood-sinus formed by the fused veins into the dorsal vessel or heart (see Lankester's 'Zoology,' part ii, introductory chapter; A. and 0. Black, 1900). The only exception to this is in the case of minute degenerate forms where the heart has disappeared altogether. The rigidity of the integument caused by the deposition of dense chitin upon it is intimately connected with the physiological activity and form of all the internal organs, and is undoubtedly correlated with the total disappearance of the circular muscular layer of the body-wall present in ChEetopods. (2) In all existing Arthropoda the 528 B. BAY LANKBSTEK. region in front of the mouth is no longer formed by the primitive prostomium or head-lobe, but one or more segments, originally post-oral, with their appendages have passed in front of the mouth (prosthonieres). At the same time the prostomium and its appendages cease to be recog- nisable as distinct elements of the head. The brain no longer consists solely of the nerve-ganglion mass proper to the pi'ostomial lobe, as in Cbasfcopoda, but is a composite (syncerebrum) produced by the fusion of this and the nerve- ganglion masses proper to the prosthomeres or segments which pass forwards, whilst their parapodia ( = appendages)

ri FIG. 1*.—Diagram to show the gradual formation of the Arthropod pericardia! blood-sinus aud "ostiate" heart by the swelling up (phlebcedesis) of the veins entering the dorsal vessel or heart of a Chtetopod-like ancestor. The figure on the left represents the condition in a Cheetopod, that on the right the condition in an Arthropod; the other two are hypothetical intermediate forms. (After Lankester, 'Quart. Journ. Micr. Sci.,' vol. xxxiv, 1893.)

become converted into eye-stalks and antennas, or more rarely grasping organs. (3) As in Chsetopoda, ccelomic funnels (ccelomoduets) may occur right and left as pairs in each ring-like segment or somite of the body, and some of these are in all cases retained as gonoducts and often as renal excretory organs (green glands, coxal glands of Arachnida—not crural glands, which are epidermal in origin); but true nephridia, genetically identical with the nephridia of earthworms, do not occur (on the subject of STBUCTCJJRB AND CLASSIFICATION OP THI5 AllTUROPODA. '529 coelom, coeloruoducts, and nephridia, see the introductory chapter of part ii of Lankester's ' Treatise on Zoology'). Tabular Statement of the Grades, Classes, and Sub-classes of the Arthropoda.—It will be convenient now to give in the clearest form a statement of the larger subdivisions of the Arthropoda which it seems necessary to recognise at the present day. The justification of the arrangement adopted will form the substance of the rest of the present article. The orders included in the varions classes are not discussed here, but are treated of under the following titles:—PEKJPATUS (Onychophora), MYRIAPODA (Diplopoda and Chilopoda), ARAUHNIDA, INSJECTA (Hexapoda), and CBUSTACEA.

STJB-PHYLUM ARTHKOPODA (of the Phylum Appeu- diculata). Grade A. Hyparthropoda (hypothetical forms connecting au- cestors of Chaetopoda with those of Arthropoda).

Grade B. Protarthropoda. Class ONYCHOFHORA. Ex.—Peripatus.

Grade C. Euarthropoda. Class 1. DJPLOPODA. Ex.—Julus. Class 2. AEACHNIDA. Grade a. Anomomeristica. Ex.— Phacops. Grade b. Nomomeristica. (a) Pantopoda. Ex.—Pycnogonum. (b) Euarachmda. Ex.—Limulus, Scorpio, Mygale, Aca- rus. 530 E. BAY LANKESTJEB.

Class 3. CRUSTACEA. Grade a.—Entoniostraca. Ex.—Apus, Branchipus, Cyclops, Balanus. Grade b. Malacostraca. Ex.—Nebalia, Astacus, Oniscus, Gam- marus. Class 4. CHILOPODA. Ex.—Scolopendra. Class 5. HBXAPODA (syn. Insecta Pterygota). Ex.—Locusta, Phryganea, Papilio, Apis, Musca, Cimex, Lucanus, Machilis. Incet'tsa sedis.—Tardigrada, Pentastomidas (degenerate forms).

The Segmentation of the Body of Arthropoda.— The body of the Arthropoda is more or less clearly divided iuto a series of rings, segments, or somites, which can be shown to be repetitions one of another, possessing identical parts and organs which may be larger or smaller, modified in shape or altogether suppressed in one somite as compared with another. A similar constitution of the body is more clearly seen in the Chsetopod worms. In the Vertebrata also a repetition of units of structure (myotomes, vertebrae, etc.) — which is essentially of the same nature as the repetition in Arthropods and Chastopods, but in many respects subject to peculiar developments—is observed. The name "metamerism " has been given to this structural phenomenon because the " meres," or repeated units, follow one another in line. Each such " mere " is often called a " metamere." This is not the place in which to discuss the origin and essential nature of " metamerism " or " metameric segmentation." Nevertheless a satisfactory consideration of the structure of the Arthropoda demands a knowledge of what may be called the laws of metamerism. These are not so fully ascertained or formulated as might be expected. The repetition of parts, which we note as metamerism, is, as Haeckel, Bateson, and others have recognised, only a special STRUCTURE AND CLASSIFICATION OP THE ARTHROPODA. 531 case of a tendency of the organic body to repetition of structural units or parts which, finds one expression in bilateral symmetry. In certain worms (the Cestoidea and some Planarians) metarneric segmentation is accompanied by the separation of the completed metameres one by one from the older (anterior) extremity of the chain (strobilation), but it by no means follows that inetameric segmentation has a necessary origin in such completion and separation of the " meres." On the contrary, metamerism seems to arise from a property of organisms which is sometimes more (eumero- genesis) and sometimes less (dysmerogenesis) fully exhibited, and in some groups not exhibited at all. Tlie most complete and, at the same time, simplest instances of metameric segmentation are to be seen in the larger Chsetopods, where some hundreds of segments succeed oue another—each practically indistinguishable in structure from the segment in front or from that behind; muscles, right and left appendage or parapodium, colour pattern of the skin, gut, blood-vessels, coelom, nephridia, nerve-ganglion, and nerves are precisely alike in neighbouring segments. The segment which is least like the others is the first, for that carries the mouth and a lobe projecting beyond it—the prostomium. If (as sometimes happens) any of the hinder segments completes itself by developing a prostomium, the chain breaks at that point, and the segment which has developed a prostomium becomes the first or head-bearing segment of a new individual. Compare such an instance of metameric segmentation with that presented by one of the higher Arthropods—e. g. the crayfish. Here the somites are not so clearly marked in the tegumentary structures; nevertheless, by examining the indications given by the paired parapodia, we find that there are twenty-one somites present—a limited definite number which is also the precise number found in all the higher Crustacea. We can state as a FIRST LAW1 of metamerism or somite formation that it is either indefinite in regard to number of 1 The word "LATT" is used in this summary merely as a convenient heading for the statement of a more or less general proposition. 532 E. BAY LANKKS'J'BR. metameres or somites produced or is definite. Animals in the first case we call anomomeristic; those in the second case, nomomeristic. The nornomeristic condition is a higher development, a. specialisation, of the anomomeristic condition. The SECOND LAW, or generalisation, as to metamerism which must be noted is that the meres or somites (excepting the first with its prostomium) may be all practically alike, or may differ from one another greatly by modification of the various constituent parts of the mere or somite. Metamerised animals are either hotncoomeric or heteromeric. The reference to the variation in the form of the essential parts contained in a " metamere " or " somite " introduces us to the necessity of a general term for these constituent or subordinate parts; they may be called " meromes" (/aipog). The meromes present in a metamere or somite differ in different annulate or segmented animals according to the general organisation of the group to which the animal belongs. As a matter of con- venience we distinguish in the Arthropod as meromes, first, the tegumentary chitinised plates called terga, placed on the dorsal aspect of the somites ; second, the similar sternal plates. In Chastopods we should take next to these the masses of circular and longitudinal muscular fibres of the body-wall and the dorso-ventral muscles. The latter form the third sort of raerome present in the Arthropods. The fourth kind of merome is constituted by the parapodia or appendages; the fifth by the coelornic pouclies and their ducts and external apertures (coelomo-ducts), whether renal or genital. The sixth by the blood-vessels of the somite; the seventh by the bit of alimentary tract which traverses it; and the eighth by the neuromere (nerve-ganglion pah', commissures, connectives, and nerve branches). It becomes apparent from this enumeration that there are a good many important elements or " meromes " in an Arthro- pod metamere or somite which can become the subject of heteromerism, or, to use a more apt word, of " heterosis." It is all the more necessary to insist upon this, inasmuch as there is a tendency in the discussion of the segmentation of the STRUCTURE AND CLASSIFICATION OF THE ARTHROPODA. 533

Arthropod body to rely exclusively upon the indications given by the tegumentary chitinous plates and the parapodia. The THIRD LAW of metamerism is that heteroraerism may operate in sucli a way as to produce definite regions of like modification of the somites and their appendages, differing in their modification from that observed in regions before and behind them. It is convenient to have a special word for such regions of like meres, and we call each a tagma (ray/ia, a regiment). The word " tagmosis" is applicable to the formation of such regions. In the Chaetopods tagmosis always occurs to a small extent, so as to form the head. In some Chtetopods, such as Chastopterus and the sedentary forms, there is marked tagmosis, giving rise to three or even more tagmata. In Arthropods, besides the head, we find very frequently other tagmata developed. But it is to be noted that in the higher members of each great class or line of descent, the tagmosis becomes definite and characteristic just as do the total number of meres or somites, whilst in the lower grades of each great class we find what may be regarded as varying examples of tentative tagmosis. The terms nomotagmic and anomotagmic are applicable with the same kind of implication as the terms nomomeristic and anomomeristic. The FOURTH LAW of metamerism (auto-heterosis of the meroines) is that the meromes of a somite or series of somites may be separately and dissimilarly affected by heteromerism. It is common enough for small changes only to occur in the inner visceral meromes, whilst the appendages and terga or sterna are largely changed in form. But of equal importance is the independent "heterosis" of these visceral meromes without any corresponding heterosis of the body-wall. As instances we may cite the gizzards of various earthworms, and the special localisation of renal, genital, and gastric meromes, with obliteration elsewhere, in a few somites in Arthropod a. The FIFTH LAW, relating also to the independence of the meromes as compared with the whole somite, is the law of autorhythmus of the meromes. Metamerism does not always 534 E. KAY LANKESTER.

manifest itself in the formation of complete new segments; but one merome may be repeated so as to suggest several metameres, whilst the remaining meromes are, so to speak, out of harmony with it and exhibit no repetition. Thus in the hinder somites of the body of Apus the Crustacean we find a series of segments corresponding apparently each to a complete single somite, but when the appendages are examined we find that they have multiplied without relation to the other meromes of a somite; we find that the somites carry from two to seven pairs of appendages, increasing in number as we pass backwards from the genital segment. The appendages are autorhythmic meromes in this case. They take on a quasi-independent metamerism, and are produced in numbers which have no relation to the numbers of the body-rings, muscles, and neuromeres. This possibility of the independent metameric multiplication of a single merome must have great importance in the case of dislocated meromes, and no doubt has application to some of the metameric phenomena of Vertebrates. A case which appears at first sight to be one of " auto- l'hythmus" of the parapodia is that of the Diplopods (Julus, etc.), in which each apparent somite carries two pairs of legs or parapodia. It looks at first as though this were due to the independent multiplication of the legs ; but it is not. Con- trary to what obtains in Apus, we find in Julus that there is a well-marked somite in the embryo corresponding to each pair of legs, and that the adult condition arises from a fusion of the tegumentary meromes of adjacent somites (see below, "Fusion"). The SIXTH LAW is the law of dislocation of meromes. This is a very important and striking phenomenon. A merome, such as a pair of appendages (Araneas) or a neuromere, or a muscular mass (frequent), may (by either a gradual or sudden process, we cannot always say which) quit the metamere to which it belongs, and in which it originated, and pass by actual physical transference to another metamere. Frequently this new position is at a distance of several metameres from STRUCTURE AND CLASSIFICATION OF THE ARTHBOPODA. 535 that to which the wandering naerome belongs in origin. The movement is more usual from behind forwards than in the reverse direction ; but this pi'obably has no profound signifi- cance, and depends simply on the fact that, as a rule, the head must be the chief region of development on account of its containing the sense organs and the mouth. In the Vertebrata the independence of the meromes is more fnlly developed than in other metamerised animals. Not only do we get auto-heterosis of the meromes on a most extensive scale, but the dislocation of single meromes and of whole series (tagmata) of meromes is a common phenomenon. Thus in fishes the pelvic fins may travel forwards to a thoracic and even jugular position in front of the pectoral fins; the branchiomeromes lose all relation to the position of the meromes of muscular, skeletal, ccelomic, and nervous nature, and the heart and its vessels may move backwards from their original metameres in higher Vertebrates carrying nerve-loops with them. The SEVENTH LAW of metamerism is one which has been pointed out to the writer by Mr. E. S. Goodrich, of Merton College, Oxford. It may be called the law of " translation of heterosis." Whilst actual physical transference of the substance of meromes undeniably takes place in such a case as the passage of the pelvic fins of some fishes to the front of the pectorals, and in the case of the backward movement of the opisthosomatic appendages of. spiders, yet the more frequent mode in which an alteration in the position of a specialised organ in the series or scale of metameres takes place is not by migration of the actual material organ from somite to somite, but by translation of the quality or morphogenetic peculiarity from somite to somite accompanied by correlative change in all the somites of the series. The phenomenon may be compared to the transposition of a piece of music to a higher or lower key. It is thus that the lateral fins of fishes move up and down the scale of vertebral somites;1 and thus that whole regions (tagmata), such as those indicated 1 Except in such cases as have just been cited.—U. R. L., 1904. 536 E. BAY LANKRSTER. by the names cervical, thoracic, lumbar, and sacral, are translated (accompanied by terminal increase or decrease in the total number of somites) so as to occupy differing numerical positions in closely allied forms (cf. the varying number of cervical somites in allied reptiles and birds). What, in this rapid enumeration, we will venture to call the EIGHTH LAW of metamerism is the law of homceosis, as it is termed by Bateson (1). Homoeosis is the making of a merome into the likeness of one belonging to another metamere, and is the opposite of the process of "heterosis"— already mentioned. We cite this law here because the result of its opei'ation is to simulate the occurrence of dislocation of meromes, and has to be carefully distinguished from that process. A merome can and does, in individual cases of abnormality, assume the form and character of the corresponding merome of a distant somite. Thus the antenna of an insect has been found to be replaced by a perfectly well- formed walking leg. After destruction of the eye-stalk of a shrimp a new growth appears, having the form of an antenna. Other cases are frequent in Crustacea as individual abnor- malities. They prove the existence in the mechanism of metamerised animals of structural conditions which are capable of giving these results. What those structural conditions are is a matter for separate inquiry, which we cannot even touch here. It is not improbable that homoeosis of distant meromes may have given rise to permanent structural changes characteristic of whole groups of Arthropoda, supposing the abnormality once established to be favoured by natural selection. Possibly the chelate condition of the prteoral appendages of Arachnida maybe due to homoeosis trans- ferring the chelate form of post-oral limbs to what were previously antonniform rami. We now come to the question of the production of new somites or the addition of new somites to the series, and the converse problem of the suppression of somites, whole or partial. We state as the NINTH LAW of metamerism " that new somites or metameres are added to a chain consisting of STRUCTURE AND CLASSIFICATION OF THE AttTHROPODA. 537 two or more somites by growth and gradual elaboration— what is called "budding"—of the anterior border of the hindermost somite. This hindermost somite is therefore different from all the other somites, and is called the ' telson.' However long or short or heteromerised the chain may be, new metameres or somites are only produced at the anterior border of the telson, except in the Vertebrata." That is the general law; but amongst some groups of metamerised animals partial exceptions to it occur. It is probably absolutely true for the Arthropoda from lowest to highest. It is not so certain that it is true for the Chastopoda, and would need modification in statement to meet the cases of fissi- parous multiplication occurring among Syllids and Naidids. In the Vertebrata, where tagmosis and heterosis of meromes and dislocation of meromes and tagmata are, so to speak, rampant, new formation of metameres (at any rate as represented by important meromes) takes place at more than one point in the chain. Such points are found where two highly diverse "tagmata" abut on one another. It is possible, though the evidence at present is entirely against the supposition, that at such points in Arthropoda new somites may be formed.1 Such new somites are said to be "intercalated." The question of the intercalation of vertebi-Ee in the Verte- brata has received some attention. It must be remembered that a vertebra, even taken with its muscular, vascular, and ueural accessories, is only a partial metamere—a merome, and that, so far as complete inefcameres are concerned, the

1 The curious case of superabundant parapodia in the binder somites of Apus has already been cited and referred to as an example of autorbythmic multiplication of meromes. There is some reason for regarding the extra pairs of legs as being "intercalated" after the formation of the somite as a single unit or merome by growth from the telson. Supposing, as appears to be the case, that as the Apus increases in size, the number of extra pairs of legs on a non-terminal somite increases, these added meromes are certainly intercalated, and represent incomplete intercalated metameres. The intercalation of new elements does not really go much further than this in Vertebrata, for a vertebra with its myoskcletal tissues is only a merome, and not a complete metamere. VOL. 47, PART 4. NEW SEEIES. M M 538 E. RAY LANKESTT5B. Vertebrata do conform to the same law as the Arthropods. Intercalation of meromes—branchial, vertebral, and dermal (fin-supports)—seems to have taken place in Vertebrata in the fishes, while in higher groups intercalation of vertebrae in large series has been accepted as the only possible explanation of the structural facts established by the comparison of allied groups. The elucidation of this matter forms a very important part of the work lying to the hand of the investi- gator of vertebrate anatomy, and it is possible that the application of Groodrich's law (the seventh of our list) may throw new light on the matter. In regard to the diminution in the number of somites in the course of the historical development of those various groups of metamerised animals, which have undoubtedly sprung from ancestors with more numerous somites than they themselves possess, it appears that we may formulate thefollowiug laws as the tenth, eleventh, twelfth, and thirteenth laws of metamerism. The TENTH LAW is that individual somites tend to atrophy and finally disappear as distinct structures, most readily at the anterior and the posterior ends of the series constituting an animal body. This is very generally exhibited in the head of Arthropoda, where, however, the operation of the law is largely modified by fusion (see below). With regard to the posterior end of the body, the atrophy of segments does not, as a rule, affect the telson itself so much as the somites in front of it and its power of producing new somites. Some- times, however, the telson is very minute and non-chitinised

The ELEVENTH LAW may be stated thus:—Any somite in the series which is the anterior or posterior somite of a tagma may become atrophied, reduced in size, or partially aborted by the suppression of some oE its meromes ; and finally, such a somite may disappear and leave no obvious trace in the adult structure of its presence in ancestral forms. This is called the excalation of a somite. Frequently, however, such "escalated" somites are obvious in the embryo or leave some STRUCTURE AND CLASSIFICATION OF THE ARTHROPODA. 539 merome (e. g. neuromere, muscle, or chitin-plate) which can be detected by minute observation (microscopic) as evidence of their former existence. The somite of the masillipede (thirdpost-oral appendage) of Apus cancrif ormis is agood example of a somite on its way to excalation. The third pra> oral and the prsemaxillary somites of Hexapod insects are instances where the only traces of the vanished somite are furnished by the microscopic study of early embryos. The prsegenital somite of the Arachnida is an example of a somite which is preserved in some members of the group and partially or entirely excalated in other cases, sometimes with fusion of its remnants to neighbouring somites. The TWELFTH LAW of metamerism might very well be placed in logical order as the first. It is the law of lipomer- isni, and asserts that just as the metameric condition is produced by a change in the bodies of the descendants of unisegmental ancestors, so highly metamerised forms, i. e. strongly segmented forms with specialised regions of differ- entiated metameres, may gradually lose their metamerised structure and become apparently and practically unisegmental animals. The change here contemplated is not the atrophy of terminal segments one by one so as to reduce the size of the animal and leave it finally as a single somite. On the contrary, no loss of size or of high organisation is necessary. But one by one, and gradually, the metameric grouping of the bodily structures disappears. The cuticle ceases to be thickened in rings; the muscles of the body-wall overrun their somite boundaries. Internal septa disappear. The nerve-ganglia concentrate or else become diffused equally along the cords; one pair of renal coelomoducts and one pair of genital coelomoducts grow to large size and remain—the rest disappear. The appendages atrophy or become limited to one or two pairs, which are widely dislocated from their ancestral position. The animal ceases to present any indication of metameric repetition of parts in its entire structure. Degrees in this process are frequently to be recognised. We certainly can observe such a change in the posterior region 540 B. BAY LANKBSTEB. of some Arthropods, such as the hermit crabs and the spiders. Admitting that the Bchiurids are descended from Chfetopoda, such a change has taken place in them amounting to little short of complete lipomerism, though not absolutely complete. Receut suggestions as to the origin of the Mollusca involve the supposition that such an effacement of once well-marked metamerism has occurred in them, leaving its traces only in a few structures such as the multiple gill-plumes and shell- shields of the Chitons and the duplicated renal sacs of Nautilus. A further matter of importance in this connection is that when the old metameres have been effaced a new secondary segmentation may arise, as in the jointed worm-like body of the degenerate Acarus, Demodex f olliculorum. Such secondary annulation of the soft body calls to mind the secondary annulation of the metameres of leeches and some earthworms. Space does not permit of more than an allusion to this subject, but it is worth while noting that the secondary annuli marking the somites of leeches and Lumbricidte in definite number and character are perhaps comparable to the redundant pairs of appendages on the hinder somites of Apus, and are in both cases examples of independent repetition of tegumentary meromes—a sort of ineffectual attempt to subdivide the somite which only pre- vails on the more readily susceptible meromes of the integument. The development of secondary metameric annulations within the area of a complete somite is not recorded among Arthropoda. It deserves distinct recognition as " hypo- metamerism " or formation of " sornatidia." The last law of metamerism which we shall attempt to formulate here, as the THIRTEENTH, relates to the fusion or blending of neighbouring somites. There are, without doubt, a large number of important generalisations to be arrived at hereafter from the further study of the metamerism of Vertebrata and the peculiar phenomena exhibited by the dislocated meromes of the vertebrate's somites. But this is STRUCTURE AND CLASSIFICATION OF THE ARTHROPODA. 541 not the place in which to attempt an outline of the special laws of vertebrate metamerism. Fusion of adjacent somites has often been erroneously interpreted in the study of Arthropoda. There are, in fact, very varying degrees of fusion which need to be carefully distinguished. The following generalisation may be formulated :—" The homologous meromes of two or more adjacent somites tend to fuse with one another by a blending of their substance. Very generally, but not invariably, the fused meromes are found as distinct separated structures in the embryo of the animal in which they unite at a later stage of growth." The .fusion of neighbouring meromes is often preceded by more or less extensive atrophy of the somites concerned, and by a.rrest of development in the individual ontogeny._, Thus a case of fusion of partially atrophied somites may simulate the ap- pearance of incipient merogeuesis or formation of new somites; and vice versa, incipient merogenesis may be misinterpreted as a case of fusion of once separate and fully formed somites. Moreover the two phenomena, merogenesis and fusion of meromes, actually occur side by side in some cases, as in the pygidial shields of the Trilobitae and Limulus. The most commonly noted cases of fusion of metameres are simply cases of the fusion of the tegumentary meromes— usually the terga only. Such a fusion has really no very serious morphological importance: it is superficial and readily acquired. It amounts to no more than the disposition of chitinous cuticle of equal thickness over the area of the terga of the somites concerned instead of the thinning of the cuticular deposit at the adjacent borders of the somites. The somites consequently lose their hinge; they can no longer be flexed one on the other. Atrophy of the muscles related to such flexure necessarily follows. The mesosomatic portion of the posterior carapace of Limulus is no more than such a superficial fusion: the other meromes of the ankylosed somites (appendages, neuromeres, blood-vessels, etc.) are unaffected. Such, too, is the case with the pygidial 542 E. BAY LANKESTEB. shields of many Trilobites. On the other hand, the telson, which is joined in both these cases with the superficially fused segments by a fusion of its chitinous cuticle with that of its last-formed or budded somite, can only take part in the fusion as a result of arrest in its activity, which amounts to a late supervening atrophy. This arrest of the telson's special bud growth may take place very early, in which case we get a large telsonic shield and only a very few somites in front of it—none soldered to the telson as in Agnostus and Ilenus; or it may take place later when eight post-cephalic (opisthosomatic) somites have been formed as in Limulus— the last two incompletely. Or, again, thirty or more somites may have been produced before the arrest takes place, and fifteen of these may be ankylosed with the telson to foi'm the pygidial shield (Phacops, etc.). A more complete fusion of somites is that seen in the head of Arthropoda. The head or prosoma of Arthropod a is a tagma consisting of one, two, or three prosthomeres or somites in front of the mouth, and of one, two, three, up to five or six opisthomeres. The cephalic tagma or prosoma may thus be more or less sharply divided into two sub- tagmata, the praa-oral and the post-oral. The shifting of the mouth backwards in Arthropoda so as to allow segments which once were post-oral to take up a prseoral position, as prosthomeres, must be regarded as a case of dislocation of the meromes concerned (sixth law), like the forward travelling of a fish's pelvic fins. The anus does not appear to be liable to such dislocation in Arthro- poda, but it certainly does travel away from its parental metamere in the Vertebrata, and may possibly do so in Cha;topoda when what must be called " lipomerism" or general obliteration of a metameric ordering of parts sets in. Such "lipomerism" must be supposed to have affected the Chastopod ancestors of the Sipunculids, if those latter worms are to be traced genetically to the former, and the anus has shifted to the anterior third of the body. However that may be, the conception (first put forward by Lankester in 1875) STfeUCTUBE ANO CLASSIFICATION OF l'HE AETHEOPODA. 543 (2) of the backward movement of the mouth in Arthropoda from the first somite to the second, third, or even fourth in the original post-oral series, is not only justified by embryological observation of the shifting in question, but finds its parallel in other instances of the law of dislocation of nieromes. The fusion of the cephalic or prosoruatic somites not only extends to tegumentary structures, but to muscles, blood- vessels, and markedly to neuromeres. However, in the embryo of many Arthropoda the original neuromeres of the praeoral somites can be distinguished, and in many cases the coslomic cavities. Also it is a noteworthy fact that the tegumentary fusion (cephalic carapace, prosomatic carapace) appears sometimes to break down secondarily (e. g. Squilla among Crustacea and Galeodes and Tarassidae among Arach- nida). It appears that we must recognise as a principle that such fusions as the carapaces of Arthropoda can revert to the condition of free movable plates; and therefore we must not assume that forms with fused tergal plates are necessarily later, genetically, than allied forms with free movable tergal plates. When such reversion to a movable series of dorsal plates occurs it must not be assumed that any coiTesponding change takes place in the deeper meromes. On the whole, fusion and ankylosis of somites is not in itself necessarily a deep- seated or far-reaching process. It may or may not be accompanied by dislocation • of important meromes or by lipomerism; whilst,—as for instance in the opisthosoma of the spiders, opiliones, and acari—dislocation and lipomerism may occur without fusion of tegumentary plates, and with, on the contrary, a dwindling and eventual atrophy of such plates. The genei-al considerations as to metamerism set forth above will enable us to proceed to a consideration of the characters which distinguish the various groups of Arthro- poda, and to justify the classification with which we started. The Theory of the Arthropod Head.—The arthropod 544 E. BAY LANKESTER. head is a tagma or group of somites which differ in number and in their relative position in regard to the mouth, in different classes. In a simple Ohastopod (fig. 1) the head consists of the first somite only; that somite is perforated by the mouth, and is provided with a prostomium or praeoral lobe. The prostomium is essentially a part or outgrowth of the first sornite, and cannot be regarded as itself a somite. It gives rise to a nerve-ganglion mass, the prostomial ganglion. In the marine Chastopods (the Polychasta) (fig. 2) we find

FIG. 1. FIG. 2.

Pr-

]?IG. 1.—Diagram of the Lead and adjacent region of an Oligo- chajte Clisetopod. Pr, tlie prostomium; «t, the mouth; A, the prostomial ganglion-mass or archicerebrum; I, II, III, ccelom of the first, second, and third somites. (From Goodrich, 'Q. J. Micr. Sci.,' vol. xi, p. 247.) FIG. 2.—Diagram of the head and adjacent region of a Polychsete Chsetopod. Letters as in Pig. 1, with the addition of T, prostomial tentacle ; Pa, parapodium. (From Goodrich.) the same essential structure, but the prostomium may give rise to two or more tactile tentacles, and to the vesicular eyes. The somites have well-marked parapodia, and the second and third, as well as the first, may give rise to tentacles which are directed forward, and thus contribute to form " the head." But the mouth remains as an inpushing of the wall of the first somite. The Arthropoda are all distinguished from the Chastopoda by the fact that the head consists of one or more somites which He in front of the mouth (now called prosthomeres), STRUCTURE AND CLASSIFICATION OP THE ARTHROPODA. 545 as well as of one or more somites behind it (opisthomeres). The first of the post-oral somites invariably has its parapodia modified so as to form a pair of hemignattis (mandibles). Twenty-five years ago the question arose as to whether the somites in front of the mouth are to be considered as derived from the prostomium of a Ohsetopod-like ancestor. Milne- Edwards and Huxley had satisfied themselves with discussing and establishing, according to the data at their command, the number of somites in the Arthropod head, but had not considered the question of the nature of the praoral somites. Lankester (2) was the first to suggest that (as is actually the fact in the Nauplius larva of Crustacea) the prseoral somites or prosthomeres and their appendages were ances- trally post-oral, but have become prsBoral " by adaptational shifting of the oral aperture." This has proved to be a sound hypothesis, and is now accepted as the basis upon which the Arthropod head must be interpreted (see Korschelt and Heider [3]). Further, the morphologists of the •'fifties appear, with few exceptions, to have accepted a preliminary scheme with regard to the Arthropod head and Arthropod segmentation generally, which was misleading aud caused them to adopt forced conclusions and interpretations. It was conceived by Huxley, among others, that the same number of cephalic somites would be found to be characteristic of all the diverse classes of Arthropoda, and that the somites not only of the head, but of the various regions of the body, could be closely compared in their numerical sequence in classes so distinct as the Hexapods, Crustaceans, and Arachnids. The view which it now appears necessary to take is, on the contrary, this—viz. that all the Arthropoda are to be traced to a common ancestor resembling a Chastopod worm, but differing from it in having lost its chastaa and in having a prosthomere in front of the mouth (instead of prostomium only) and a pair of hemignaths (mandibles) on the parapodia of the buccal somite. From this ancestor Arthropods with heads of varying degrees of complexity have been developed 546 a. RAY LANKESTM. characteristic of the different classes, whilst the parapodia and somites of the body have become variously modified and grouped in these different classes. The resemblances which the members of one class often present to the members of another class in regard to the form of the limb-branches (rami) of the parapodia, and the formation of tagmata (regions) are not hastily to be ascribed to common inherit- ance, but we must consider whether they are not due to homoplasy—that is, to the moulding of natural selection acting in the different classes upon fairly similar elements under like exigencies. The structure of the head in Arthropods presents three profoundly separated grades of structure dependent upon the number of prosthomeres which have been assimilated by the prteoral region. The classes presenting these distinct plans of head-structure cannot be closely associated in any scheme of classification professing to be natural. Peripatus, the type genus of the class Onycliophora, stands at the base of the series with only a single prosthomere (fig. 8). In Peripatus the prostoinium of the Chtetopod-like ancestor is atrophied, but it is possible that two processes on the front of the head (FP) represent in the embryo the dwindled prostornial tentacles. The single prosthomere carries the retractile tentacles as its "parapodia." The second somite is the buccal somite (II, fig. 3); its parapodia have horny jaws on their ends, like the claws on the following legs (fig. 8), and act as hemignaths (mandibles). The study of sections of the embryo establishes these facts beyond doubt. It also shows us that the neuromeres, no less than the embryonic coelomic cavities, point to the existence of one, and only one, prosthomere in Peripatus, of which the " Protocerebrum," P, is the neuromere, whilst the Deutero- cerebrum, D, is the neuromere of the second or buccal somite. A brief indication of these facts is given by saying that the Onycliophora are " deuterognathous,"—that is to say, that the buccal somite carrying the mandibular hemignaths is the second of the whole series. STRUCTURE AND CLASSIFICATION 01' THE ARTHROPODA. 647 What has become of the nerve-ganglion of the prostoinial lobe of the Chtetopod in Peripatus is not clearly ascertained, nor is its fate indicated by the study of the embryonic head of other Arthropods so far. Probably it is fused with the protocerebruin, and may also be concerned in the history of the vei'y peculiar paired eyes of Peripatus, which are like those of Chastopods in structure—viz. vesicles with an intra-

FlG. 3.—Diagram of the liead and adjacent region of Peripatus. Monoprostliomerous. in, Mouth; I, coelom of the first somite which carries the antennae, and is in front of the mouth; II, coelom of the second somite, which carries the mandibles (hence deuterognathous); III and IV, coelom of the third and fourth somites; F.P., rudimentary frontal processes, perhaps representing the prostomial tentacles of Poljchseta; Ant, antenna or tactile tentacle; Md, mandible; Op, oral papilla; P, prot.oceiebrum or foremost cerebral mass belonging to the first somite; D, deuterocerebrum, consisting of ganglion cells belonging to the second or mandibular somite. (After Goodrich.) vesicular lens, whereas the eyes of all other Arthropods have essentially another structure, being "cups" of the epidermis, in which a knob-like or rod-like thickening of the cuticle is fitted as refractive medium. In Diplopoda (Julus, etc.) the results of enibryological study point to a composition of the front part of the head exactly similar to that which we find in Onychophora. They are deuterognathous. The Arachnida present the first stage of progress. Here embryology shows that there are two prosthomeres (fig. 4), 548 B. KAY LANKESTEE. and that the guatkobases of the chelas which act as the first pair of heinignaths, are carried by the third somite. The Araclmida are therefore tritoguathous. The two prosthomeres are indicated by their cceloinic cavities in the embryo (I and II, fig. 4), and by two neuromeres, the protocerebrum and the deuterocerebrum. The appendages of the first prosthomere are not present as tentacles, as in Peripatus and Diplopods, but are possibly represented by the eyes or possibly altogether aborted. The appendages of the second prosthomere are the well-known chelicerte of the Arachnids, rarely,

FIG. 4.—Diagram of the head and adjacent region of an Arachnid. Diprostbomerous in the adult condition, though embryologically the appendages of somite II and the somite itself are, as here drawn, not actually in front of the mouth. E, lateral eye ; Ch, chelicera ; HI, mouth; P, piotocerebrum ; D, deuterocerebium ; I, II, III, IV, ccelom of the first, second, third, and fourth somites. (After Goodrich.) if ever, antenniform, but modified as " retroverts" or clasp- knife fangs iu spiders. The Crustacea (fig. 5) and the Hexapoda (fig. 6) agree in having three somites in front of the mouth, and it is probable, though not ascertained, that the Chilopoda (Scolopendra, etc.) are in the same case. The three prosthomeres or prteoral somites of Crustacea due to the sinking back of the, mouth one somite farther than in Araclmida are not clearly indicated by ccelomic cavities in the embryo, but their existence is clearly established by the development and position of the appendages and by the neuromeres. STRUCTURE AND CLASSIFICATION OF THE ARTHBOPODA. 549 The eyes in some Crustacea are mounted on articulated stalks, and from the fact that they can after injury be replaced by antenna-like appendages it is inferred that they represent the parapodia of the most anterior prosthomere. The second prosthomere carries the first pair of antennas and

FIG. 5. FIG. 6.

TIG. 5.—Diagram of the head of a crustacean. Triproslho- nierous. F. P, frontal processes (observed in Cinhiped nauplius- larvae), probably representing the prostomifil tentacles of Chaetopods; e, eye ; Ant1, first pair of antennae ; Ant2, second pair of antennae; md, mandidle; ma;1, imp, first and second pairs of maxillse; m, mouth; I, II, III, tlie three prosUiomeres; IV, V, VI, the three somites following the mouth; P, proctocerebrum ; D, deuterocerebrum; T, tritocerebrum. (After Goodrich.) FIG. 6.—Diagram of the head of a Hexapod insect, e, eye; ant, antenna; md, mandible; ma;1, first maxilla; ma2, second maxilla; in, mouth; I, region of the first or eye-bearing piosthomere; II, coelom of the second antenna-bearing prosthomere; III, coelom of the third prosthomere devoid of appendages; IV, V, and VI, coelom of the fourth, fifth, and sixth somites ; P, protocerebrum belonging to the first, prosthomere ; D, deutei'Ocerebrum belonging to the second prosthomere; T, tritocerebrum belonging to the third prosthomere. (After Goodrich.) the third the second pair of antennae. Sometimes this pair of appendages has not a merely tactile jointed ramus, but is converted into a claw or clasper. Three neuromeres—a proto-, deutero-, and tritocerebrum—corresponding to those three prosthomeres, are sharply marked in the embryo. The fourth somite is that in which the mouth now opens, and which accordingly has its appendages converted into 550 E. KAY LANKESTEE. hemignafchous mandibles. The Crustacea are tetartognathous. The history of the development of the head has been carefully worked out in the Hexapod insects. As in Crustacea and Arachnida, a first prosthomere is indicated by the paired eyes and the protocerebrum; the second prosthomere has a well-marked coelomic cavity, carries the antennae, and has the deuterocerebrurn for its neuroniere. The third prosthomere is represented by a well-marked pair of coelomic cavities and the tritocerebrum (III, fig. 6), but has no appendages. They appear to have aborted. The existence of this third prosthomere, corresponding to the third prosthomere of the Crustacea, is a strong argument for the derivation of the Hexapoda, and with them the Chilopoda, from some offshoot of the Crustacean stem or class. The buccal somite, with its mandibles, is in Hexapoda, as in Crustacea, the fourth: they are tetartognathous. The adhesion of a greater or less number of somites to the buccal somite posteriorly (opisfchomeres) is a matter of importance, but of minor importance, in the theory and history of the Arthropod head. In Peripatus no such adhesion or fusion occurs. In Diplopoda two opisthomeres—that is to say, one in addition to the buccal somite—are united by a fusion of their terga with the terga of the prosthomeres. Their appendages are respectively the mandibles and the gnathochilar i a m. In Arachnida the highest forms exhibit a fusion of the tergites of five post-oral somites to form one continuous carapace united with the terga of the two prosthomeres. The five pairs of appendages of the post-oral somites of the head or prosoma thus constituted all primitively carry gnathobasic projections on their coxal joints, which act as hemignaths; in the more specialised forms the mandibular gnathobases cease to develop. In Crustacea the fourth or mandibular somite never has less than the two following somites associated with it by the adaptation of their appendages as jaws, and the ankylosis of STRUCT UK E AND CLASSIFICATION OF THE ARTHROPODA. 551 their terga with that of the prosthomeres. But in higher Crustacea the cephalic "tagma" is extended, and more somites are added to the fusion, and their appendages adapted as jaws of a kind. The Hexapoda are not known to us in their earlier or more primitive manifestations; we only know them as possessed of a definite number of somites arranged in definite numbers in three great tagmata. The head shows two jaw-bearing somites besides the mandibnlar somite (V, VI, in fig. 6) — thus six in all (as in some Crustacea), including prosthomeres, all ankylosed by their terga to form a cephalic shield. There is, however, good embryological evidence in some Hexapods of the existence of a seventh somite, the supra-lingual, occurring between the somite of the mandibles and the somite of the first maxillas (4). This segment is indicated einbryo- logically by its paired coelomic cavities. It is practically an excalated somite, having no existence in the adult. It is probably not a mere coincidence that the Hexapod, with its two rudimentary somites devoid of appendages, is thus found to possess twenty-one somites, including that which carries the anus, and that this is also the number present in the Mala- costracous Crustacea. The Segmental Lateral Appendages or Limbs of Arthropoda.—It has taken some time to obtain any general acceptance of the view that the parapodia of the Cheetopoda and the limbs of Arthropoda are genetically identical structures ; yet if we compare the parapodium of Tomopteris or of Phyllodoce with one of the foliaceons limbs of Branchipus or Apus the correspondences of the two are striking. An erroneous view of the fundamental morphology of the crustacean limb, and consequently of that of other Arthropoda, came into favour owing to the acceptance of the highly modified limbs of Astacus as typical. Protopodite, endopodite, exopodite, and epipodite were considered to be the morphological units of the crustacean limb. Lankester (5) has shown (and his views have been accepted by Professors Korschelt and Heider in their treatise on 'Embi'yology') 552 B. JUT LANKESTER. that the limb of the lowest Crustacea, such as Apus, consists of a corm or axis which may be jointed, and gives rise to out-

FIG. 7. FIG. 8.

nt.c.

PIG. 7.—Diagram of the somite-appendage or parapodiuni of a Polyclisete Clisetopod. The choetrc are omitted. Ax, the axis; nr. c, neuropodial cirrhus; nr. I1, nr. I1, neuropodial lobes or endites; nt. c, notopodial cirrhus ; nt. I1, nt. P, notopodial lobes or exites. The parapodium is represented with its neural or ventral surface uppermost. (Original.) PIG. 8.—Three somite-appendages or parapodia of Peripatus. A, a walking leg ; p to ;;', the characteristic "pads;"/", the foot; el1, cP, the two claws; B, an oral papilla, one of the second pair of post-oral appendages; C, one of the first post-oral pair of appendages or mandibles; cl\ cP, the greatly enlarged claws. (Compare A.) The appendages are represented with the neural or ventral surface uppermost. (Original.) growths, either leaf-like or filiform, on its inner and outer margins (endites and exites). Such a corm (see figs. 9 and 10), with its outgrowths, may be compared to the simple STRUCTURE AND CLASSIFICATION OF THE ARTHROPODA. 553 parapodia of Chtetopoda with cirrhi and branchial lobe (fig. 7). It is by the specialisation of two "eudites" thab the endopodite and exopodite of higher Crustacea are formed, whilst a flabelliform exite is the homogen or genetic equivalent of the epipodite (see Lankester, " Observations and Reflections on Apus cancriformis/' fQ. J. Micr. Sci.'). The reduction of the outgrowth-bearing " corm" of the parapodium of either a Chtetopod or an Arthropod to a simple cylindrical stump, devoid of outgrowths, is brought about

FIG. 9.—The second thoracic (fifth post-oral) appendage of the left side of Apus cancriformis, placed with its ventral or neural surface uppermost to compare with Figs. 7 and 8. 1, 2, the two segments of the axis; en\ t,he gnathobase ; en- to en6, the five following "endites;" fl, the flahellum or anterior exite; ir, the bract or"posterior exite. (After Lankester, ' Q. J. Micr. Sci.,' vol. xxi, 1881.) when mechanical conditions favour such a shape. We see it in certain Chsetopods (e. g. Hesione) and in the Arthropod Peripatus (fig. 8). The conversion of the Arthropod's limb into a jaw, as a rule, is effected by the development of an endite near its base into a hard, chitinised, and often toothed gnathobase (see figs. 9 and 10, en1). It is not true that all the biting processes of the Arthropod limb are thus produced, —for instance, the jaws of Peripatus are formed by the axis or corm itself; whilst the poison-jaws of Chilopods, as also their maxillas, appear to be formed rather by the apex or terminal region of the ramus of the limb; but the opposing jaws ( = hemignaths) of Crustacea, Arachnida, and Hexapoda are gnathobases, and not the axis or corm.. The endopodite (corresponding to the fifth endite of the limb of Apus, see fig. 9) becomes in Crustacea the " walking leg " of the mid- VOL. 47, PART 4.—NKW SEKTES. UN ' 554 E. RAY LANKBSTKR. region of the body ; it becomes the palp or jointed process of anterior segments. A second ramus, the "exopodite," often is also retained in the form of a palp or feeler. In Apus, as

FIG. 10.—The first thoracic (fourth post-oral) appendage of Apus cancriformis (right side). Axx to A-v*, the four segments of the axis with muscular bands; En\ gnathobase ; En1 to En?, the elongated jointed endites (rami); En6, the rudimentary sixth endite (exopodite of higher Crustacea); Fl, the flabellum which becomes the epipodite of higher forms; Br, the bract devoid of muscles and respiratory in function. (AJter Laukester, 'Q. J. Micr. Sci.,' vol. xxi, 1881.)' the figure shows, there are four of these " antenna-like" palps or filaments on the first thoracic limb. A common modification of the chief ramus of the Arthropod parapodium is the chela or nipper formed by the elongation of the pen-

1 This figure lias been re-drawn for the present reprint.—E, R. L. STEUCTTJRE AND CLASSIFICATION OP THIS AETHEOPODA. 555

ultimate joint of the ranius, so that the last joint works on it —as, for instance, in the lobster's claw. Such chelate rami or limb-branches are independently developed in Crustacea and in Arachnida, and are carried by somites of the body

rvf FIG. 11.—Diagram to show the derivation of the unit or " ommatidium " of the compound eye of Crustacea and Hexapoda, C, from a simple monomeniscous monostichous eye resembling the lateral eye of a scorpion, A, or the unit of the compound lateral eye of Limulus (see article AUACHNIDA, Figs. 22 and 23). B represents an intermediate hypothetical form in which the cells beneath the lens are beginning to be superimposed as corneagen, vitrella, and retinula, instead of standing side by side in horizontal series. The black represents the cuticular product of the epidermal cells of the ocular area, taking the form either of lens, cl, of crystalline body, cry, or of rhabdom, rhab; hy, hypodermis or epidermal cells; corn, laterally placed cells in the simpler stage A, which like the nerve-end cells, vil1 and ret1, are corneagens or lens-producing; corn, specialised corneagen or lens-producing cells ; vil1, potential vitrella cells with cry, potential crystalline body now indistinguishable from retinula cells and rhabdomeres ; vil, vitrella cell with cry, its contained cuticular product, the crystalline cone or body ; vil1, rhab1, retinula cells and rhabdom of scorpion undifferentiated from adjacent cells, vil1; ret1, retinula cell; rhab, rhabdom ; nf, optic nerve-fibres. (Modified from Watasae.)

which do not correspond in position in the two groups. The range of modification of which .the rami or limb-branches of the limbs of Arthropoda are capable is veiy large, and in allied orders, or even families or genera, we often find what is certainly the palp of the same appendage (as determined 556 H. RAY LANKESTER. by numerical position of the segments)—in one case antenniform, in another chelate, in another pediform, and in another reduced to a mere stump or absent altogether. Very probably the power which the appendage of a given segment has of assuming the perfected form and proportions previously attained by the appendage of another segment must be classed as an instance of " homoeosis," not only where such a change is obviously due to abnormal development or injury, but also where it constitutes a difference permanently established between allied orders or smaller groups, or between the two sexes. The most extreme disguise assumed by the Arthropod parapodium or appendage is that of becoming a mere stalk supporting an eye, a fact which did not obtain general credence until the experiments of Herbst, in 1895, who found, on cutting off the eye-stalk of Patemon, that a jointed antenna-like appendage was regenerated in its place. Since the eye-stalks of Podothalmate Crustacea represent appendages, we are forced to the conclusion that the sessile eyes of other Crustacea, and of other Arthropoda generally, indicate the position of appendages which have atrophied.1 From what has been said it is apparent that we cannot, in attempting to discover the affinities and divergences of the various forms of Arthropoda, attach a very high phylogenetic value to the coincidence or divergence in form of tbe appendages belonging to the somites compared with one another. The principal forms assumed by the Arthropod parapodium and its rami may be thus enumerated : (1) Axial corm well developed, unsegmented or with two to four segments ; lateral endites and exites (vami) numerous and of various lengths (certain limbs of lower Crustacea). (2) Corm, with short, unsegmented rami, forming a flat- 1 H. Milne-Edivards, who was followed by Huxley, long ago formulated the conclusion that the eye-stalks of Crustacea are modified appendages, basing his argument on a specimen of Palinmus (figured in Bateson's book) (1), in which the eye-stalk of one side is replaced by an antenniform palp. Hofer (6) in 1894 described a similar case in Astacus. STBUCTURU AND CLASSIFICATION OF THE AETHBOPODA. 557 tened foliaceous appendage, adapted to swimming and respira- tion (trunk limbs of Phyllopods). (3) Corm alone developed, with no endites or exites, but provided with terminal chitinous claws (ordinary leg of Peripatus), with terminal jaw teeth (jaw of Peripatus), or with blunt extremity (oral papilla of same) (see fig. 8). (4) Three of the raini of the primitive limb (endites 5 and 6 and exite 1) specially developed as endopodite, exopodite, and epipodibe, the first two often as firm and strongly chitinised, segmented, leg-like structures; the original axis or corm reduced to a basal piece, with or without a distinct gnathobase (endite 1), typical trirarnose limb of higher Crustacea. (5) One ramus (the endopodite) alone developed—the original axis or corm serving as its basal joint with or without gnathobase. This is the usual uniraruose limb found in the various classes of Arthropoda. It varies as to the presence or absence of the jaw process and as to the stoutness of the segments of the ramus, their number (frequently six,plus the basal corm),and the modification of the free end. This may be filiform, or brush-like, or lamellate when it is an antenna or palp; a simple spike (walking leg of Crustacea, of other aquatic forms, and of Chilopods and Diplopods); the terminal joint flattened (swimming leg of Crustacea and Gigantostraca); the terminal joint provided with two or with three recurved claws (walking leg of many terrestrial forms—e.g. Hexapoda and Arachnida); the penultimate joint with a process equal in length to the last joint, so as to form a nipping organ (chelas of Crusta- ceans and Arachnids); the last joint reflected and movable on the penultimate, as the blade of a clasp-knife on its handle (the retrovert, toothed so as to act as a biting jaw in the Hexapod Mantis, the Crustacean Squilla, and others; with the last joint produced into a needle-like stabbing process in spiders). (6) Two rami developed (usually, but perhaps not always, the equivalents of the endopodite and exopodite) supported on the somewhat elongated corm (basal segmeno). This is 558 B. RAY 1ANKBSTEB. the typical "biramose limb" often found in Crustacea. The rami may be flattened for swimming, when it is " a biramose switnmeret," or both or only one may be filiform and finely annulate; this is the form often presented by the antennas of Crustacea, and rarely by piaaoral appendages in other Arthropods. (7) The endopoditic ramus is greatly enlarged and flattened, without or with only one jointing, the conn (basal segment) is evanescent; often the plate-like endopodites of a pair of such appendages unite in the middle line with one another or by the intermediary of a sternal upgrowth and form a single broad plate. (These are the plate-like switn- merets and opercula of Grigantostraca and Limulus among Arachnids and of Isopod crustaceans. They may have rudimentary exopodites, and may or may not have branchial filaments or lamella) developed on their posterior faces. The simplest form to which they may be reduced is seen in the genital operculum of the scorpion.) (8) The gnathobase becomes greatly enlarged and not separated by a joint from the corm; ib acts as a hemignath or half-jaw working against its fellow of the opposite side. The endopodite may be retained as a small segmented palp at the side of the gnathobase or disappear (mandible of Crustacea, Chilopoda, and Hexapoda). (9) The corm becomes the seat of a development of a special visual organ, the Arthropod eye (as opposed to the Chfetopod eye). Its jointing (segmentation) may be retained, but its rami disappear (podophthalmous Crustacea). Usually it becomes atrophied, leaving the eye as a sessile organ upon the prseoral region of the body. (The eye-stalk and sessile lateral eyes of Arthropoda generally, exclusive of Peripatus.) (10) The forms assumed by special modification of the elements of the parapodium iu the maxillee, labium, etc., of Hexapods, Chilopods, Diplopods, and of various Crustacea deserve special enumeration, but cannot be dealt with without ample space and illustration. STRUCTURE AND CLASSIFICATION OF TfiE AETHBOPODA. 559 It may be pointed out that the most radical difference presented in this list is that between appendages consisting of the conn alone without rarni (Onychophora) and those with more or less developed rami (the rest of the Arthropoda). In the latter class we should distinguish three phases: (a) those with numerous and comparatively undeveloped rarni; (b) those with three, or two highly developed rami, or with only one—the corrn being reduced to the dimensions of a mere basal segment; (c) those reduced to a secondary simplicity (degeneration) by overwhelming development of one segment (e. g. the isolated gnathobase often seen as "mandible" and the genital opercnlum). There is no reason to suppose that any of the forms of limb observed in Arthropoda may not have been independently developed in two or more separate diverging lines of descent. Branchiae.—In connection with the discussion of the limbs of Arthropods a few words should be devoted to the gill- processes. It seems probable that there are branchial plumes or filaments in some Arthropoda (some Crustacea) which can be identified with the distinct branchial organs of Chastopoda, which lie dorsad of the parapodia and are not part of the parapodium. On the other hand, we cannot refuse to admit that any of the processes of an Arthropod parapodium may become modified as branchial organs, and that, as a rule, branchial outgrowths are easily developed, de novo, in alJ the higher groups of animals. Therefore it seems to be, with our present knowledge, a hopeless task to analyse the branchial organs of Arthropoda and to identify them genetically in groups. A brief notice must suffice of the structure and history of the Byes, the Tracheas, and the so-called Malpighian tubes of Arthropoda, though special importance attaches to each in regard to the determination of the affinities of the various animals included in this great sub-phylum. The Eyes.—The Arthropod eye appears to be an organ of special character developed in the common ancestor of the 560 B. EAT LANKESTER. Euarthropoda, and distinct from the Chastopod eye, which is found only in the Onychophora where the true Arthropod eye is absent. The essential difference between these two kinds of eye appears to be that the Chaatopod eye (in its higher developments) is a vesicle enclosing the lens, whereas the Arthropod eye is a pit or series of pits into which the heavy chitinous cuticle dips and enlarges knobwise as a lens. Two distinct forms of the Arthropod eye are observed—the monomeniscous (simple) and the polymeniscous (compound). The nerve-end cells, which lie below the lens, are part of the general epidermis. They show in the monomeniscous eye (see article ARACHNIDA,1 fig. 26) a tendency to- group themselves into " retinulse," consisting of five to twelve cells united by vertical deposits of chitin (rhabdoms). In the case of the polymeniscous eye (fig. 23, article ARACHNIDA) a single retinula or group of nerve-end cells is grouped beneath each associated lens. A further complication occurs in each of these two classes of eye. The monomeniscous eye is rarely provided with a single layer of cells beneath its lens; when it is so, it is called monostichous (simple lateral eye of scorpion, fig. 22, article ARACHNIDA). More usually, by an infolding of the layer of cells in development, we get three layers under the lens; the front layer is the corneagen layer, and is separated by a membrane from the other two, which more or less fuse and contain the nerve-end cells (retinal layer). These eyes are called diplostichous, aud occur in Arachnida aud Hexapoda (fig. 24, in article ARACHNIDA.). On the other hand, the polymeniscous eye undergoes special elaboration on its lines. The retinulas become elongated as deep and very narrow pits (fig. 11 and explanation), and develop additional cells near the mouth of the narrow pit. Those nearest to the lens are the corneagen cells of this more elaborated eye, and those between the original retinula cells and the corneagen cells become firm and. transparent. They are the crystalline cells or vitrella (see Watase, 7). 1 Tuis article will be reproduced from the ' Encyclopedia' in the next number of this Journal,—E. It. L. STRUCTURE AND CLASSIFICATION OF THE ARTHROPODA. 561 Each such complex of cells underlying the lenticle of a compound eye is called an " ommatidium;" the entire mass of cells underlying a monomeniscous eye is an "ommataaum." The ommatasum, as already stated, tends to segregate into retinulse which correspond potentially each to an ommatidium of the compound eye. The ommatidium is from the first segregate, and consists of few cells. The compound eye of the king-crab (Limulus) is the only recognised instance of ommatidia in their simplest state. Bach can be readily compared with the single-layered lateral eye of the scorpion. In Crustacea and Hexapoda of all grades we find compound eyes with the more complicated ommatidia described above- We do not find them in any Arachnida. It is difficult, in the absence of more detailed knowledge as to the eyes of Chilopoda and Diplopoda, to give full value to these facts in tracing the affinities of the various classes of Arthropods. But they seem to point to a community of origin of Hexapods and Crustacea in regard to the complicated ommatidia of the compound eye, and to a certain isolation of the Arachnida, which are, however, traceable, so far as the eyes are concerned, to a distant common origin with Crustacea and Hexapoda through the very simple compound eyes (monostichous, polymeuiscous) of Limulus. The Trachete.—In regard to tracheae the very natural tendency of zoologists has been until lately to consider them as having once developed and once only, and therefore to hold that a group " Tracheata " should be recognised, including all tracheate Arthropods. We are driven by the conclusions arrived at as to the derivation of the Arachnida from branchiate ancestors, independently of the other tracheate Arthropods (see article ARACHNIDA), to formulate the conclusion that trachete have been independently developed in the Arachnidan class. We are also, by the isolation of Peripatus and the impossibility of tracing to it all other tracheate Arthropoda, or of regarding it as a degenerate offset from some one of the tracheate classes, forced to the conclusion that the tracheae of the Onychophora have been independently acquired. Having 562 E. RAY LANKESTEft. accepted these two conclusions, we formulate the generalisation that tracheae can be independently acquired by various branches of Arthropod descent in adaptation to a terrestrial as opposed to an aquatic mode of life. A great point of interest, therefore, exists in the knowledge of the structure and embryology of tracheae in the different groups. It must be confessed that we have not such full knowledge on this head as could be wished for. Tracheae are essentially tubes like blood-vessels—apparently formed from the same tissue elements as blood-vessels—which contain air in place of blood, and usually communicate by definite orifices, the trachea! stigmata, with the atmosphere. They are lined internally by a cuticular deposit of chitin. In Peripatus and the Diplopods they consist of bunches of fine tubes which do not branch, but diverge from one another; thechitinouslining is smooth. In the Hexapods and Ohilopods, and the Arachnids (usually), they form tree-like branching structures, and their finest branches are finer than any blood capillary, actually in some cases penetrating a single cell and supplying it with gaseous oxygen. In these forms the chitinous lining of the tubes is thickened by a close-set spiral ridge similar to the spiral thickening of the cellulose wall of the spiral vessels of plants. It is a noteworthy fact that other tubes in these same terrestrial Arthropoda—namely, the ducts of glands—are similarly strengthened by a chitinous cuticle, and that a spiral or annular thickening of the cuticle is developed in them also. Ohitin is not exclusively an ectodermal product, but occurs also in cartilaginous skeletal plates of mesoblastic origin (connective tissue). The immediate cavities or pits into which the tracheal stigmata open appear to be in many cases ecto- dermic in sinkings, but there seems to be no reason (based on embryological observation) for regarding the tracheae as an ingrowth of the ectoderm. They appeal', in fact, to be an air-holding modification of the vasifactive connective tissue. Tracheae are abundant just in proportion as blood-vessels become suppressed. They are reciprocally exclusive. It seems not improbable that they are two modifications of the STBUCTUBE AND CLASSIFICATION OF THE ARTHROPODA. 563 same tissue elements. In Peripatus the stigmatic pits at which the tracheae communicate with the atmosphere are scattered and not definite in their position. In other cases the stigmata are definitely paired and placed in a few segments or in several. It seems that we have to suppose that the vasifactive tissue of Arthropoda can readily take the form of air-holding instead of blood-holding tubes, and that this somewhat startling change in its character has taken place independently in several instances—viz. in the Onychophora, in more than one group of Arachuida, in Diplopoda, and, again, in the Hexapoda and Chilopoda. The Malpighian Tubes.—This name is applied to the numerous fine csecal tubes of noticeable length developed from the proctodasal invert of ectodermal origin in Hexapods. These tubes are shown to excrete nitrogenous waste products similar to uric acid. Tubes of renal excretory function in a like position occur in most terrestrial Arthropoda—viz. in Chilopoda, Diplopoda, and Arachnida. They are also found in some of the semi-terrestrial and purely aquatic Amphipod Crustaceans. But the conclusion that all such tubes are identical in essential character seems to be without founda- tion. The Malpighian tubes of Hexapods are outgrowths of the proctodteum, but those of scorpion and the Amphipod Crustacea are part of the meteuteron or endodermal gut, though originating near its junction with the proctodseum. Hence the presence or absence of such tubes cannot be used as an argument as to affinity without some discrimination. The scorpion's so-called Malpighian tubes are not the same organs as those so named in the other Tracheata. Such i-enal caacal tubes seem to be readily evolved from either mefcenteron or proctodseum when the conditions of the outwash of nitrogenous waste products are changed by the transference from aquatic to terrestrial life. The absence of such renal casca in Limulus and their presence in the terrestrial Arachnida is precisely on a parallel with their absence in aquatic Crustacea and their presence in the feebly branchiate Amphipoda. We shall now pass the groups of the Arthropoda in review, 564 E. RAY LANKBSTEK. attempting to characterise them in such away as will indicate their probable affinities aud genetic history. SUB-PHYLUM ARTHROPODA.—The characters of the sub- phylum, and those of the associated, sub-phyla Cheetopoda and Rotifera, have been given above, as well as the general characters of the phylum Appendiculata which comprises these great sub-phyla. Grade A.—Hyparthropoda. Hypothetical forms. Grade B.—Protarthropoda. (a) The integument is covered, by a delicate soft cuticle (not firm or plated) which allows the body and. its appendages great range of extension and contraction. (b) The paired claws on the ends of the parapodia and the fang-like modifications of these on the first post-oral append- dages (mandibles) are the only hard chitinous portions of the integument. (c) The head is deuterognathous,—that is to say, there is only one prosthomere, and accordingly the first and only pair of hemignaths is developed by adaptation of the appendages of the second somite. (d) The appendages of the third somite (second post-oral) are clawless oral papillse. (e) The rest of the somites carry equi-formal simple appendages, consisting of a coim or axis tipped with two chitinous claws and devoid of rami. (/) The segmentation of the body is anomeristic, there being no fixed number of somites characterising all the forms included. (g) The pair of eyes situated on the prosthomere are not of the Euarthropod type, but resemble those of Chsetopods (hence Nereid- ophthalmous). (h) The muscles of the body-wall and gut do not consist of transversely striped muscular fibre, but of the unstriped tissue observed also in Chastopoda. (i) A pair of ccelomoducts is developed in every somite, STRUCTURE AND CLASSIFICATION OF THE ARTHROPODA. 565 including the prosthomere, in which alone it atrophies in later development. (j) The ventral nerve-cords are widely separated,—in fact, lateral in position. (k) There are no masses of nerve-cells forming a ganglion (neuromere) in each somite. (In this respect the Protarthro- poda are at a lower stage than most of the existing Chaeto- poda.) (I) The genital ducts are formed by the enlargement of the coalomoducts of the penultimate somite.

Class (Unica).—ONTCHOPHOBA. With the characters of the grade : add the presence within the body oE fine unbranched trachea! tabes, devoid of spiral thickening, opening to the exterior by numerous irregularly scattered trachea! pits. Genera—Eoperipatus, Peripatopsis, Opisthopatus, etc.

Grade C (of the Arthropoda).—Euarthropoda. (a) Integument heavily plated with firm chitinous cuticle, allowing no expansion and retraction of regions of the body nor change of dimensions, except, in some cases, a dorso- ventral bellows movement. The separation of the heavier plates of chitin by grooves of delicate cuticle results in the hinging or jointing of the body and its appendages, and the consequent flexing and extending of the jointed pieces. (b) Claws and fangs are developed on the branches or rami of the parapodia, not on the end of the axis or corm. (c) The head is either deuterognathous, tritognathous, or tetartognathous. (d) Rarely only one, and usually at least two, of the somites following the mandibular somite carry appendages modified as jaws (with exceptions of a secondary origin). (e) The rest of the somites may all carry appendages, or only a limited number may carry appendages. In all cases the appendages primarily develop rami or branches 566 E. RAY LANKESTEB.

which form the limbs, the primitive axis or corm being reduced and or' insignificant size. In the most primitive stock all the post-oral appendages had gnathobasic outgrowths. (/) The segmentation of the body is anomomeristic in the more archaic members of each class, nomomeristic in the higher members. (g) The two eyes of Chsetopod structure have disappeared, and are replaced by the Euarthropod eyes. (h) The muscles in all parts of the body consist of striped muscular fibre, never of unstriped muscular tissue. (i) The coelomoducts are suppressed in most somites, and retained only as the single pair of genital ducts (very rarely more numerous), and in some also as the excretory glands (one or two pairs). (j) The ventral nerve-cords approach one another in the mid-ventral line behind the mouth. (k) The nerve cells of the ventral nerve-cords are segregated as paired ganglia in each somite, often united by meristic dislocation into composite ganglia. (1) The genital ducts may be the ccelomoducts of the penultimate or antepenultimate or adjacent somite, or of a somite placed near the middle of the series, or of a somite far forward in the series.

Class 1 (of the Euarthropoda).—DIPLOPODA. The head has but one prosthomere (monoprosthomerous), and is accordingly deuterognathons. This carries short-jointed antennas (in one case biramose) and eyes, the structure and development of which require further elucidation. Only one somite following the first post-oral or mandibular segment has its appendages modified as jaws. The somites of the body, except in Pauropus, either fuse after early development and form double somites with two pairs of appendages (Julus, etc.) or present legless and leg- bearing somites alternating. STRUCTURE AND CLASSIFICATION OF THE ARTHROPODA. 567

Somites, anomomeristic, from 12 to 150 in the post-cephalic series. The genital ducts open in the fourth, or between the fourth and fifth post-oral somite. Terrestrial forms with small-jointed legs formed by adaptation of a single ramus of the appendage. Tracheas are present. • Note.—The Diplopoda include the Juliformia, the Sym- phyla (Scolopendrella), and Pauropoda (Pauropus). They were until recently classified with the Chilopoda (centipedes), with which they have no close affinity, but only a superficial resemblance. (Compare the definition of the class Chilopoda.) The movement of the legs in Diplopoda is like that of those of Peripatus, of the Phyllopod Crustacea, and of the parapodia of Chsetopoda, symmetrical and identical on the two sides of the body. The legs of Chilopoda move in alternating groups on the two sides of the body ; this implies a very much higher development of nerves and muscles in that group.1

Class 2 (of the Euarthropoda).—ARACHNIDA. Head tritognathous and diprosthomerous,—that is to say, with two prosthomeres; the first bearing typical eyes, the second a pair of appendages reduced to a single ramus, which is in more primitive forms antenniform, in higher forms chelate or retrovert. The ancestral stock was pantognatho- basic, i. e. had a gnathobase or jaw process on every parapodium. As many as six pairs of appendages following the mouth may have an enlarged gnathobase actually functional as a jaw or hemignath, but a ramus is well developed on each of these appendages either as a simple walking leg, a palp, or a chela. In the more primitive forms the appendage of every post-oral somite has a gnathobase and two rami; in higher specialised forms the guathobases maybe atrophied in every appendage, even in the first post-oral.

1 See the Appendix ab the end of the present article, and the accompanying plate, 568 E. KAY LANKESTER.

The more primitive forms are anomomeristic; the higher forms nomomeristic, showing typically three groups or tagmata of six somites each. The genital apertures are placed on the first somite of the second tagma or mesosoma. Their position is unknown in the more primitive forms. The more primitive forms have branchial respiratory processes developed on a ramus of each of the post-oral appendages. Iu higher specialised forms these branchial processes become first of all limited to five segments of the mesosoma, then sunk beneath the surface as pulmonary organs, and finally atrophied, their place being taken by a well-developed tracheal system. A character of great diagnostic value in the more primitive Arachnida is the tendency of the chitinous investment of the tergal surface of the telson to unite during growth with that of the free somites in front of it, so as to form a pygidial shield or posterior carapace, often comprising as many as fifteen somites (Trilobites, Limulus). A pair of central monomeniscous diplostichous eyes is often present on the head. Lateral eyes also are often present, which are monostichous with aggregated lenses (Limulus) or with isolated lenses (Scorpio), or are diplostichous with simple lens (Pedipalpi, Aranete, etc.).

Class 3 (of the Euarthropoda).—CRUSTACEA. Head tetartognathous and triprosthomei'ous,—that is to say, with three prosthomeres : the first bearing typical eyes, the second a pair of antenniform appendages (often biramose), the third a pair of appendages, usually antenniform, sometimes claw-like. The ancestral stock was (as in the Arach- nida) pantoguathobasic,—that is to say, had a gnathobase or jaw-process on the base of every post-oral appendage. Besides the first post-oral or mandibular pair, at least two succeeding pairs of appendages are modified as jaws. These have small and insignificant rami, or none at all,—a feature in which the Arachnida differ from them. The appendages of STJUJCTUUE AND CLASSIFICATION OF THE AETHEOPODA. 569 four or more additional following somites may be turned upwards towards the mouth and assist in the taking of food. The more primitive forms (Entomosti'aca) are anomomeristic, presenting great variety as to number of somites., form of appendages, and tagmatic grouping; the higher forms (M.ala- costraca) are nomomeristic, showing in front of the telsou twenty somites, of which the six hinder carry swimmerets, and the five next in front ambulatory limbs. The genital apertures are neither far forward nor far backward in the series of somites, e. g. on the fourteenth post-oral in Apus, on the ninth post-oral in female Astacns and in Cyclops. With rare exceptions, branchial plates are developed either by modification of a ramus of the limbs or as processes on a ramus, or upon the sides of the body. No tracheate Crus- tacea are known, but some terrestrial Isopoda develop pulmonary in-sinkings of the integument. A characteristic comparable in value to that presented by the pygidial shield of Arachuida is the frequent development of a pair of long appendages by the penultimate somite, which, with the telson, form a trih'd, or when that is small a bifid termination to the body. The lateral eyes of Crustacea are poly meniscous, with highly specialised retinulas like those of Hexapoda, and unlike the simpler compound lateral eyes of lower Arachnida. Mono- meniscous eyes are rarely present, and when present single, minute, and central in position. Note.—The Crustacea exhibit a longer and more complete series of forms than any other class of Arthropoda, and may be regarded as preserving the most completely represented line of descent.

Class 4.—CHILOPODA. Head triprosthomerous' and tetartognathous. The two somites following the mandibular or first post-oral or buccal 1 Embryological evidence of this is still wantiug. In the other classes of Arthropoda we have more or less complete embryological evidence on the subject. It appears from observation of the embryo that whilst the first VOL. 47, PART 4. NEW SERIES. 0 0 570 B. BAY LANKBSTER. somite carry appendages modilied as niaxillas. The fourth post-oral somite has its appendages converted into very large and powerful hernignaths, which are provided with poison- glands. The remaining somites carry single-clawed walking legs, a single pair to each somite. The body is anomomeristic, showing in different genera from 17 (inclusive of the anal and genital) to 175 somites behind that which bears the poison-jaws. No tagmata are developed. The genital ducts opeu on the penultimate somite. Trachese are developed which are dendriform and with spiral thickening of their lining. Their trunks open at paired stigmata placed laterally in each somite of the trunk or in alternate somites. Usually the tracheae open by paired stigmata placed upon the sides of a greater or less number of the somites, but never quite regularly on alternating somites. At most they are preseDt on all the pedigerous somites excepting the first and the last. In Scutigera there are seven unpaired dorsal stigmata, each leading into a sac, whence a number of air-holding tubes project into the pericardial blood-sinus. Eenal CEecal tubes (Malpighian tubes) open into the proctodseum.

Class 5.—HEXAPODA. Head shown by its early development to be triprosthonie- rous, and consequently tetartognathous. The first prosthomere has its appendages represented by the compound eyes and a protocerebrum; the second has the antennas for its appendages and a deutocerebral neuromere; the third has suffered suppression of its appendages (which corresponded to the second pair of antennas of Crustacea), but has a tritocerebrum and coelomic chamber. The mandibular somite prosthomere of centipedes lias ils appendages reduced and represented only by eye-patches (as in Arachnida, Crustacea, and Hexapoda), the second lias a rudimentary antenna, which disappears, whilst the third carries the permanent antennee, which accordingly correspond to the second antennae of Crustacea, and are absent in Hexapoda, STRUCTURE AND CLASSIFICATION 01? TfilS AKTHROrODA. 571 bears a pair of gnathobasic hemignaths without rami or palps, and is followed by two jaw-bearing somites (maxillary and labial). This enumeration would give six somites in all to the head; three prosthomeres and three opisthomeres. Eecent investigations (Folsom, 4) show the existence in the embryo of a prsemaxillary or supra-lingual somite which is suppressed during development. This gives seven somites to the Hexapod's head, the tergites of which are fused to form a cephalic carapace or box. The number is significant, since it agrees with that found in Bdriophthalmous Crustacea, and assigns the labium of the Hexapod to the same somite numerically as that which carries the labium-like maxilli- pedes of those Crustacea. The somites following the head are strictly nomomeristic and nomotagmic. The first three form the thorax, the appendages of which are the walking legs, tipped with paired claws or ungues. (Compare the homoplastic claws of Scorpio and Peripatus.) Eleven somites follow these, forming the abdominal " tagma," giving thus twenty-one somites in all (as in the higher Crustacea). The somites of the abdomen all may carry rudimentary appendages in the embryo, and some of the hinder somites may retain their appendages in a modified form in adult life. Terminal telescoping of the abdominal somites and excalation may occur in the adult, reducing the obvious abdominal somites to as few as eight. The genital apertures are median, and placed far back in the series of somites, viz. the female on the seventh abdominal (seventeenth of the whole series) and the male on the ninth or antepenultimate abdominal (nineteenth of the whole series). The appendages of the eighth and tenth abdominal somites are modified as gonapophyses. The eleventh abdominal segment is the telson, usually small and soft; it carries the anus. The Hexapoda are not only all confined to a very definite disposition of the somites, appendages, and apertures as thus indicated, but in other characters also they present the specialisation of a narrowly limited, highly developed order 572 B. RAY IJANKESTER. of such a class as the Crustacea rather than a range from lower more generalised to higher more specialised forms such as that group and also the Arachnida present. It seems to be a legitimate conclusion that the most primitive Hexapoda were provided with wings, and that the term Pterygota might be used as a synonym of Hexapoda. Many Hexapoda have lost either one pair or both pairs of wings; cases are common of wingless genera allied to ordinary Pterygote genera. Some Hexapods which are very primitive in other respects happen to be also apterous, but this cannot be held to prove that the possession of wings is not a primitive character of Hexapods (compare the case of the Struthious birds). The wings of Hexapoda are lateral expansions of the terga of the second and third thoracic somites. They appear to be serial equivalents (homogeneous meromes) of the tracheal gills, which develop in a like position on the abdominal segments of some aquatic Hexapods. The Hexapoda are all provided with a highly developed tracheal system, which presents considerable variation in regard to its stigmata or oritices of communication with the exterior. In some a serial arrangement of stigmata comparable to that observed in Chilopoda is found. In other cases (some larvaa) stigmata are absent; in other cases again a single stigma is developed, as in the smaller Arachnida and Ghilopoda, in the median dorsal line or other unexpected position. When the facile tendency of Arthropodato develop tracheal air-tubes is admitted, it becomes probable that the tracheae of Hexapods do not all belong to one original system, but may be accounted for by new developments within the group. Whether the primitive tracheal system of Hexapoda was a closed one or open by serial stigmata in every somite remains at present doubtful, but the intimate relation of the system to the wings and tracheal gills cannot be overlooked. The lateral eyes of Hexapoda, like those of Crustacea, belong to the most specialised type of " compound eye," found only in these two classes. Simple nionomeniscous eyes are also present in many Hexapods. STRUCTURE AND CLASSIFICATION OP THE ARTHROPODA. 573 Renal excretory ceeca (Malpighian tubes) are developed from the prootodseum (not from mesenteron, as in scorpion and Amphipoda). Concluding Remarks on the Relationships to one another of the Classes of the Arthropoda.—Out- general conclusion from a survey of the Arthropoda amounts to this, that whilst Peripatus, the Diplopoda, and the Arach- nida represent terrestrial offshoots from successive lower grades of primitive aquatic Arthropoda which are extinct, the Crustacea alone present a fairly full series of representa- tives leading upwards from unspecialised forms. The latter were not very far removed from the aquatic ancestors (Trilobifces) of the Arachnida, but differed essentially from them by the higher specialisation of the head. We can gather no indication of the forefathers of the Hexapoda or of the Chilopoda less specialised than they are, whilst possessing the essential characteristics of these classes. Neither embryology nor palaeontology assists us in this direction. On the other hand, the facts that the Hexapoda and the Chilopoda have triprosthomerous heads, that the Hexapoda have the same total number of somites as the nomomeristic Crustacea, and the same number of opisthomeres in the head as the more terrestrial Crustacea, together with the same adaptation of the form of important appendages in corresponding somites, and that the compound eyes of both Crustacea and Hexapoda are extremely specialised and elaborate in structure and identical in that structure, all lead to the suggestion that the Hexapoda, and with them, at no distant point, the Chilopoda, have branched off from the Crustacean main stem as specialised terrestrial lines of descent. And it seems probable that in the case of the Hexapoda, at any rate, the point of departure was subsequent to the attainment of the nomomeristic character presented by the higher grade of Crustacea. It is, on the whole, desirable to recognise such affinities in our schemes of classification. We may tabulate the facts as to head-structure in Chastopoda and Arthropoda as follows: 574 E. RAY LANKBSTEli.

Grade x (below the Arthropoda).—AGNATHA APEOSTHOHEEA. Without parapodial jaws; without the addition of originally post-oral somites to the pi-seoral region, which is a simple prostomial lobe of the first somite ; the first somite is perforated by the mouth, and its parapodia are not modified as jaws. = CH.ETOPODA.

Grade 1 (of the Arthropoda).—MONOGNATHA MONOPEOS- THOMEEA. With a single pair of parapodial jaws carried by the somite which is perforated by the mouth; this is not the first somite, but the second. The first somite has become a prosthomere, and carries a pair of extensile antennas. = ONYCHOPHORA (Peripatus, etc.).

Grade 2 (of the Arthropoda).—DTGNATFA MONOPEOSTHOMEEA. The third somite, as well as the second, develops a pair of parapodial jaws; the first somite is a prosthomere carrying jointed antenna. = DIPLOPODA.

Grade 3 (of the Arthropoda).—PANTOGNATHA DIPEOSTHOMERA. A gnathobase is developed (in the primitive stock) on every pair of post-oral appendages ; two prosthomeres present, the second somite, as well as the first, having passed in front of the mouth, but only the second has appendages. = AEACHNIDA.

Grade 4 (of the Arthropoda).—PANTOGNATHA TEIPEOSTHOMERA. The original stock, like that of the last grade, has a gnathobase on every post-oral appendage, but three prosthomeres are now present, in consequence of the movement of the oral aperture from the third to the fourth somite. The lateral eyes are polymeniscous, with specialised vitrellse and retinulas of a definite type peculiar to this grade. = CEUSTACEA, OHILO- PODA, HEXAPODA. STRUCTURE AND CLASSIFICATION OP THE ARTHROPODA. 575 According to older views the increase of the number of somites in front of the mouth would have been regarded as a case of intercalation by new somite-budding of new prseoral somites in the series. We are prohibited by a general consideration of metamerism in the Arthropoda (see a previous section of this article) from adopting the hypothesis of intercalation of somites. However strange it may seem, we have to suppose that one by one in the course of long historical evolution somites have passed forwards and the mouth has passed backwards. In fact, we have to suppose that the actual somite which in grades 1 and 2 bore the mandibles lost those mandibles, developed their rami as tactile organs, and came to occupy a position iu front of the mouth, whilst its previous jaw-bearing function was taken up by the next somite in order, into which the oral aperture had passed. A similar history must have been slowly brought about when this second mandibnlate somite in its turn became agnathous and passed in front of the mouth. The mandibular parapodia may be supposed during the successive stages of this history to have had, from the first, well-developed rami (one or two) of a palp-like form, so that the change required when the mouth passed away from them would merely consist in the suppression of the gnathobase. The solid palpless mandible such as we now see in some Arthropoda is, necessarily, a late specialisation. Moreover it appears probable that the first somite never had its parapodia modified as jaws, but became a prosthomere with tactile appendages before parapodial jaws were developed at all, or rather pari passu with their development on the second somite. It is worth while bearing in mind a second possibility as to the history of the prosthomeres, viz. that the buccal gnathobasic parapodia (the mandibles) were in each of the three grades of prosthomerism only developed after the recession of the mouth aud the addition of one, of two, or of three post-oral somites to the praeoral region had taken place. In fact, we may imagine that the characteristic adaptation of one or more pairs of post-oral parapodia to the purposes of the mouth as jaws did 576 K. RAY LANKESTER. not occur until after ancestral forms with one, with two, and with three prosthotneres had come into existence. On the whole the facts seem to be against this supposition, though we need not suppose that the gnathobase was very large or the rarni undeveloped in the bnccal parapodia which were destined to lose their maudibular features and pass in front of the mouth.

EEPEEENCJKS. 1. BATESON.—'Materials for the Study of Variation' (Macmillan, 1894), p. 85. 2. LANKESTEE.—"Primitive Cell-layers of the Embryo," 'Annals and Mag. Nat. Hist.,' 1873, p. 336. 3. KORSCHELT and HEIDEK.—' Entwickelungsgeschichte ' (Jena, 1892), cap. xv, p. 389. 4. FOLSOM.—" Development of the Mouth Parts of Anurida," ' Bulletin Mus. Comp. Zool. Harvard College,' vol. xxxvi, No. 5, 1900, pp. 142—146. 5. TjANKTiSTisa.—"Observations and Reflections on the Appendages and Nervous System of Apus cancriformis," 'Quart. Journ. Micr.Soc.,' vol. xxi, 1881. 6. Hoi'Eit.—"Ein Krebs mit einer extremitiit statt ernes Stielauges," ' Yer- handl. d. Deutschen Zool. Gesellsch.,' 1894. 7. WATASE.—"On the Morphology of the Compound Eyes of Arthropods," 'Studies from the Biol. Lab. of the Johns Hopkins University,' vol. iv, pp. 287—334. 8. BENHAM describes backward shifting of the oral aperture in certain Chreto- pods, 'Proc. Zoolog. Soc. London,' 1900, No. lxiv, p. 976. N.B.—References to the early literature concerning the group Arthropoda will be found in Carus, 'Geschichte der Zoologie.' The more important literature up to 1892 is given in the admirable treatise on Embryology by Professors Korsclielt and Heider. STRUCTURE AND CLASSIFICATION OP THE ARTBROPODA. 577

APPENDIX.

ON THE MOVEMENTS OF THE PAEAPODIA OP PERIPATUS, MILLI- PEDES, AND CENTIPEDES.

[Matter not contained in the article published in the ' Encyclopaedia Britannica.']

I TAKE the opportunity of the issue of my article ' Arthro- poda} as a reprint to add to it some drawings showing the movement of the parapodia or legs oE Onychophora, Diplopoda, and Chilopoda. I was unable to introduce these into the original article, and I now give them in the form of a plate (PI. 42). They were made nearly twenty years ago in my laboratory at University College, London, from living specimens by Miss Stone. The live Peripatus (P. capensis) were given to me by Mr. Adam Sedgwick; the Centipede, Seolo- pendra subspinipes (Leach), was brought to me from Barbadoes by Mr. Tracey; and the Millipede, Archispiros- treptus pyrocephalus (L. Koch), I obtained from Mr. Pocock. Of course, the attempt to fix and record, by the simple use of eye and pencil, a phase of successional movement, such as that exhibited by the series of legs of the Arthropoda or of the Chtetopoda, is not altogether satisfactory at the present day. We ought to have records of these phases taken by photography and the instantaneous illumination of the electric spark. But in the meanwhile the drawings, which were carefully and conscientiously made after repeated observation and study, show some interesting facts. A fact which the drawings are not fitted to show is, that the change of phase-—that is to say, the alteration in the 578 li. RAY LANKESTEJi. angle formed by the limb and the long axis of the body, appears to proceed from behind forwards in all three cases. Each leg may be considered as resting normally at right angles to the axis of the body. Each is capable of a certain forward swing in the horizontal plane, being provided with a joint and muscles at its base, and of a corresponding backward swing in which the leg passes its first position (that perpendicular to the axis of the body) and makes an excursion ov deflection away from the perpendicular in the posterior direction. When the animal is in a state of locomotor activity all the legs steadily swing forwards and backwards through their extreme range of angular displace- ment, each at the same rate. But they do not all simul- taneously assume the same angular position relatively to the axis of the body, nor, on the other hand, do they swing irregularly. They pass consecutively from behind forwards into an identical phase of the swing movement, the leg- in front taking up the angular phase just previously exhibited by the leg behind it, which in the meantime has continued its swinging movement, either becoming more deflected or now commencing the return movement. The rate of swing is such thab in all cases as yet observed not one great wave occurs but a series of waves are produced, as when wind blows over a cornfield. These waves vary in the number of units (legs) involved in a complete wave according to the kind of Arthropod or Chsetopod under observation. The number of units involved in a "wave" or "swing-group" seems to be fixed in a given species, and not to vary according to circumstances. Whether the rate or relative rate of forward swing is always the same as that of the backward swing (which is that portion of the swing effective in pro- pulsion) has yet to be ascertained, as also the exact excursion made on each side of the perpendicular. Also it would be interesting to ascertain what are the limits of increase and diminution of the rate of swing, and what nervous mechanism, if any, is concerned in its regulation. These phenomena can only be studied satisfactorily by STRUCTURE AND CLASSIFfOATION OP THE ARTflltOPODA. 579 photography, aud require also the consideration of a large number of forms, such as a representative series of marine Chastopoda, several genera of Diplopoda, and of Chilopoda, the Phyllopod crustaceans and the higher forms, Hexapod insects and larvee. The most important fact which the drawings here published show is that in Peripatus and the Millipede the limbs on opposite sides of the body, which are morphologically related as " pairs," are always in the same phase of fore-and-aft swing; they move together and identically. On the other hand in the Centipede the pairs or opposite limbs on a segment are in phases, which are the extreme opposites in the series of positions through which the limb swings. Further, it is to be noted in connection with this that the strongly chitinised body of the Millipede takes no part by serpentine movement in the locomotory process; it remains perfectly straight. So, too, the soft body of Peripatus— though it is frequently bent and turned on itself, and may be more or less elongated and contracted at various intervals, yet does not contribute by any serpentine "stroke" to the process of locomotion. On the other hand the Centipede's locomotion is very largely effected by a powerful lateral undulation of the body—groups of three segments being alternately slightly tilted by muscular contraction first on one side and then on the other. In the case o£ the Centipede, as already noted, this serpeutine rhythmic movement of the body is accompanied by an opposition in the phase of the swing movements of those legs which are paired with one another in a single segment, and a special kind of leg and body movement is the result, with which the special forms of leg-rhythm producing locomotion in other highly-developed Arthropoda (including the tripod action in Hexapoda) might be compared with a view to a mechanical explanation of their genesis. On the other hand it is worth calling to mind that in some of the large marine CliEetopoda, viz. in Nephthys and Nereis (very few observations on the subject have been recorded) 580 E. KAT LANKESTER. the process of locomotion (when it takes the form of swimming) is very definitely assisted by a powerful serpentine movement of the whole body left and right, whilst the parapodia exhibit a very rapid (far more rapid than in terrestrial walking Arthropods) swinging action, the phases of which are identical in the paired appendages of either side of a segment, and not antagonistic in spite of the lateral undulation of the body. One of the important features in the swinging movement of tbe parapodia of Artbropoda and ChEetopoda, which can be observed by simple inspection of the living animal in movement, is the fact that the number of pairs of parapodia involved in a "swing-group" or (as we may put it) the number which one must pass in tracing the phases of movement before one comes to a pair of parapodia in exactly the same phase as that of the pair from which one started, varies in different genera and species. Sometimes the groups may be represented by a, b, c, d, e, f, g, h, a1, bl, c1, dl, e1,/1, g1, h1, a3, fc2, c2, d2, e2, f, g2, h?, where the letters of the alphabet indicate a parapodium in a given phase of swing, and a in the first group is identical in phase with a1 in the second, with as in the third, and so on. In other cases the groups are represented by two units only—a, b, a1, bl, as, b2, and so on. In P. capensis (PI. 42, fig. 4) the swing-group number is only two, a, b. The anterior unit a swings forward, whilst the posterior unit & has its claws grasping the surface, and is swinging backwards. As soon as parapodium a approaches parapodium bl (and similarly throughout the series) the movement changes, a grasps the surface, and bl (and all the others corresponding to it, viz. b, &3, bz, b*, &5, be) lets go and commences to swing forward. This is shown in the figures 4, 5, and 6 of PI. 42. In the Millipede Archispii'ostreptus, on the other hand, the swing-group number is sixteen, and (as our figures 1 and 2 of Plate 42 show) there are eight of these groups, allowing for peculiarities in the extreme anterior and posterior somites. The regions indicated by the lettering a to /in the STRUCTURE AND CLASSIFICATION OF THE AUTHROPODA. 581 figure are regions where parapodia exhibit the extreme forward swing-phase. They may be called " group-crests." Group-crests are but " phases " in the swinging of the limbs, and they pass along the whole series from behind forward, like the crest of a wave passing along a liquid. Eacli pair of successive parapodia is in turn the seat of the group-crest, and the waves keep flowing from behind forward with beautiful regularity. The rate should be measured in different forms, and the conditions affecting the rate of this rhythmic movement should be studied experimentally. In the Centipede (PI. 42, fig. 3) the " swing-group" number appears to be six, and the whole phenomenon is profoundly modified by the fact that lateral undulations of the body itself are a definite part of the locomotor activity, whilst the limbs on opposite sides of the same segment are not identical, but antagonistic in phase. It seems to me probable that the condition presented by the Centipede is a much higher development than that seen in the Millipede, and implies a unilateral differentiation of muscles and nerves which is far from primitive. It may, I think, be reckoned as one of the characters tending to separate the Diplopoda or Prosthogoiiopora altogether from association with the Chilopods. It would, of course, be very interesting in this connection to have some reliable photo- graphic studies of the phases of parapodial swing in such forms as Scutigera, and, indeed, in all families of Chilopoda.

EXPLANATION OP PLATE 42, Illustrating Professor Lankester's article on the Arthropoda. PIG. 1.—Lateral view of a specimen of Archispirostreptus pyro- cepbalus (de Kocli) drawn from a living specimen in movement. Magnified twice linear. 582 B. l(AY tiA

FIG. 2.—View of the ventral surface of the same specimen crawling on a glass plate and reflected in a mirror. The letters a to h indicate the " group- crests " or extreme phases of forward movement, which traverse the series at intervals of sixteen parapodia. FIG. 3.—Dorsal view of a living specimen of Scolopendra subspinipes (Leach) to show the lateral undulation of the body in locomotion, and the grouping of the limbs or parapodia in sixes, which are in antagonistic phases on the two sides of the same segments, but identical with those on the opposite side of the next half-group, a, c, e being in the same phase as b, d,f. PIGS. &—9.—Drawings from live specimens of Peripatus capensis to show the alternate phases of swing of the parapodia of the same side, and the identity of the phase of the right and left pairs of one and the same segment, also to show the soft-walled nature of the body, its pliability, and considerable powers of extension and contraction.

I should be glad were any of my readers able to inform me as to the name of the author of the following appreciative lines on the subject above discussed. " A centipede was happy ! Till One day a toad in fun Said,' Pray which leg Moves after which ?' This raised her doubts to such a pitch, She fell exhausted in the ditch, Not knowing how to run." Slhttri Jau/rn Jhm Set, lf»L 4- 7,KSM 4

9 Fig 1- Archispirosireptus.

;!]••); 4 ! ? 1

Fig 2.

Hulh. Lilk' I.ondoi